JP2009534664A - System and method for displaying cellular abnormalities - Google Patents

Abstract

In accordance with the principles of the present invention, methods, systems, and computer-readable media for processing image data representing cell analysis result data are provided, the processing comprising accessing image data and density based on the image data. Generating a histogram; converting the density histogram into a stepped image; performing a normalized cross-correlation between the stepped image and the reference image to measure similarity; and Determining anomaly based on similarity.

Description

  The present invention relates generally to systems and methods for processing and displaying data, and more specifically to systems and methods for processing and displaying cell analysis result data and template data in a single image.

  When analyzing the results of a target sample cell analyzer, the physician needs to compare the results of the target sample with the results of the template and also be able to analyze any anomalies in the target sample . Traditional cell analyzers provide unprocessed graphical results display in 1D, 2D, and 3D displays, using raw cell analysis results data to display only the target sample To do. The doctor analyzing the cell analysis result must physically compare the target sample result image with the template image while confirming the graphic result. These template images can be found in textbooks or in separate images. Alternatively, the doctor can keep a picture of the template image in his memory. In any case, the physician must obtain and compare these two separate images. This can be difficult because the images may not be the same scale, the same display format, etc. This makes the process of analyzing the target sample data inconvenient, inaccurate, time consuming and intensive. In addition, the resulting data from the cell analyzer is unprocessed and includes noisy, non-smooth data.

  Therefore, the data of the cell analysis result of the target sample and the template data are combined into one image in such a way that the doctor or user can accurately and efficiently analyze the target sample data in order to identify the abnormality. There is a need for systems and methods that are provided for display.

  In accordance with the principles of the present invention, as embodied and broadly described herein, methods and systems consistent with the principles of the present invention provide for the processing of image data representing data of cell analysis results; It includes steps of accessing image data, generating a density histogram based on the image data, converting the density histogram to a stepped image, and a stepped image and a reference image to measure similarity. Performing a normalized cross-correlation between and determining anomalies based on the measured similarity.

  The file of this patent or patent application contains at least one color drawing. Copies of this patent or patent application publication with color drawings will be provided by the organization upon request and payment of the necessary fee.

  The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several embodiments of the present invention and, together with the description, serve to explain the principles of the invention and together with the description illustrate features and aspects of the invention. explain.

  Reference will now be made in detail to the present invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.

(Overview)
Methods and systems consistent with the principles of some embodiments of the present invention provide a system for accessing target sample data representing cell analysis result data. As the data is accessed, the system processes the data and compares the processed target sample data to the template data. Further, the system can measure the similarity between the processed target sample data and the template data. The measured similarity can be a score format that identifies whether the target sample is normal or abnormal. Furthermore, the abnormal pattern can be flagged based on the score.

(Cell analysis)
The present invention can be used to analyze various types of cells, cell components, body fluids and / or body fluid components. The present invention is particularly useful in analyzing blood samples that contain both fluid components (serum) and solid components (various types of cells). In particular, the present invention aims to analyze cellular components in blood samples that are either whole blood (containing various types of blood cells) or fractions of cellular components. The invention can also be used to analyze cells obtained from a tissue sample separated from connective tissue and suspended in a biocompatible liquid medium that does not destroy the cells. Furthermore, the present invention can also be applied to analyze multidimensional cell or particle scatter plots obtained using conventional blood or flow cytometry instruments. The terms “cell analyzer” and “cell analysis” are intended to cover at least all of the components described herein. Further, when described with respect to target sample data, the term is intended to include data of cell analysis results of the target sample.

(Generation of raw data)
Body fluids and / or cellular components of body fluids and / or whole blood can be subjected to various types of analytical techniques to generate data for analysis and display according to the present invention. The most common techniques are the direct current method for measuring cell size volume, the radio frequency method for measuring cell opacity, the fluorescence method, and the light scattering method for measuring cell particle size. .

(Target sample data and template data)
Target sample data and / or template data can be white blood cells (WBC), red blood cells (RBC), platelets, one-dimensional histograms with complete blood count (CBC), two-dimensional and / or three-dimensional WBC differential scatter plots, two-dimensional and / or Or a three-dimensional reticulocyte differential scatter diagram, a two-dimensional or three-dimensional nucleated red blood cell (NRBC) differential scatter diagram, a WBC difference histogram of a curved surface display image, a reticulocyte difference histogram of a curved surface display image, an NRBC difference histogram of a curved surface display image , Etc., can be shown in the form of image data. As an alternative, the stored template data can be stored after the raw data has been applied to image smoothing and stepwise image conversion.

(System architecture)
FIG. 1 is an exemplary schematic diagram of a system environment 100 for implementing the principles of the present invention. The components of system 100 may be implemented by any suitable combination of hardware, software, and / or firmware. As shown in FIG. 1, the system 100 includes a user computer 102. User computer 102 may be communicatively coupled to database 104. As an alternative, the database 104 may reside directly on the network 106 or the contents of the database 104 may reside directly on the computer 102 or server 108.

  System 100 may further include a network 106 that may be implemented as the Internet or any public or private local or wide area network. The system 100 can further include a server 108 that can be communicatively coupled to the analyzer 110. The analyzer 110 may be implemented as a Beckman Coulter hematometer, such as LH750® and LH500®, and generates test result data.

  Although only one computer 102, database 104, network 106, server 108, and analyzer 110 are depicted, two or more of these types of devices are the principles of some embodiments of the present invention. It can be understood by one skilled in the art that it can be implemented in a system consistent with. It can further be appreciated that each of these devices may be at a different location in the system. For example, the analyzer 110 may be communicatively coupled to the computer 102, and the computer 102 may receive data directly from the analyzer 110 without operating over a network. Features consistent with the principles of the present invention may be implemented alone in computer 102 as a stand-alone unit, and all data required to carry out the present invention may reside directly on computer 102 and the analyzer It can further be appreciated that target sample data from 110 can be input by a user through an external device of computer 102.

  FIG. 2 shows an exemplary block diagram of components included within computer 102. The computer 102 may be any type of computing device such as a personal computer, workstation, or personal computing device, such as a memory 202, a network interface application 204, an input / output device 206, a central processing unit 208, application software. 210 and secondary storage 212 may be included. Computer 102 may be communicatively coupled to database 104, server computer 108 and / or analyzer 110.

  A user may access network 106 using network interface application 204 and / or application software 210. If network 106 may be implemented as the Internet, network interface application 204 may include a conventional browser, including a conventional browser application available from Microsoft or Netscape. Application software 210 may include programming instructions for implementing features of the invention described herein. Application software 210 may include programming instructions to allow target sample data to be displayed along with template data so that a user can review and / or analyze test result data. Input / output device 206 may include, for example, a keyboard, mouse, video cam, display, storage device, printer, and the like.

  FIG. 3 shows an exemplary block diagram of components included within server computer 108. Server computer 108 includes memory 302, network interface application 304, input / output device 306, central processing unit 308, application software 310, and secondary storage 312 consistent with the principles of some embodiments of the invention. obtain. The components of server computer 108 may be implemented similarly to the components of computer 102.

(Functionality)
FIG. 4 shows an exemplary flow diagram of the steps performed by computer 102, consistent with some embodiments of the present invention. As shown in FIG. 4, once the target sample data has been identified for analysis by the user, the computer 102 accesses the target sample data via the application software 310 (step 402). The target sample data may be data representing the result of cell analysis performed by the analyzer 110. This data may be stored on the computer 102, the database 102, or the server 108. The system then generates a concentration histogram based on the accessed target sample data (step 404). A density histogram can be generated by processing raw image data from a cell analyzer using a filter, eg, a low pass filter, to remove noise and smooth the image. A density compensation function can then be obtained and pixel values are equalized to improve appearance.

  The density histogram is then converted to a stepped image by the system (step 406). To generate a stepped image, multiple levels, for example a four level threshold, are performed to obtain a stepped image. Using the graded image, the system performs a normalized cross-correlation between the graded image and the reference image or template data to measure similarity (step 408). This can be performed using a technique based on Fast Fourier Transform (FFT). Compared with conventional cross-correlation algorithms, the FFT-based method is more computationally efficient, especially when the data size is large. Template data represents standard data against which target samples are compared. The template data can represent, for example, the average of many samples, the average of many samples from which extraneous data has been deleted, and the like. Template data may be stored on the computer 102, the database 102, or the server 108. The measured similarity can be shown in the form of a score, and if the score is high, the target sample is normal. If the score is low or falls below a predetermined threshold, the target sample is abnormal (step 410).

  FIG. 7 shows an exemplary flowchart of the steps performed by client computer 104 in determining correlation. The client computer 104 accesses the sample (step 702). After the target sample is accessed, the client computer 104 performs a normal template match to determine whether the target sample is normal (step 704). If the matching score between the normal template and the target sample exceeds a certain threshold (step 706, “Yes”), the computer 104 determines that the target sample is normal (step 708).

  If the matching score between the target sample and the normal template falls below a predetermined threshold, an abnormal template matching is executed (709). During anomaly template matching, the target sample is correlated with at least one anomaly template to identify anomalies. For example, the target sample is checked against the abnormal template 1 (step 710). If the matching score exceeds a predetermined threshold (step 712, “Yes”), the target sample is determined to have type 1 anomaly (step 714). This process can be repeated for multiple anomaly templates. If the matching score does not exceed a predetermined threshold for any of the abnormal templates to be verified, it is determined that the target sample has an unknown abnormality (step 716).

  FIG. 5 (a)-(b) shows the relationship between the original histogram data from the cell analysis result data and the processed cell analysis result data consistent with the principles of some embodiments of the present invention. An example of the difference. As shown in FIGS. 5 (a) to 5 (b), the original histogram image is shown in FIG. 5 (a), and the transformed stepped image is shown in FIG. 5 (b). Images I and II in FIG. 5 (a) both represent a normal pattern. However, the concentration of each population is different. An image III in FIG. 5A is a sample having an abnormal pattern. The cross-correlation coefficient used to measure the similarity between the two images is shown next to the arrow. FIG. 5 (b) shows a result based on a transformed stepped image consistent with the principles of some embodiments of the present invention. As can be seen, the similarity measure between the two normal samples (images I and II) is increasing. In contrast, the similarity measure between normal and abnormal samples (Image III) is significantly reduced.

  As can be seen from FIG. 5 (a) and FIG. 5 (b), the stepwise image conversion can compensate for the difference in density of the original image. Instead of using raw histogram images for template matching, processing cell analysis results data in the manner described herein provides clearer information between normal and abnormal patterns. Can be provided. Furthermore, each level of the stepped image is a binary image. Thus, the user can easily obtain a lot of useful image information, eg, the number of populations at a given level based on the analysis of binary images.

  After processing the cell analysis result data as described above, the system may display the data to the user. For example, the system may display the transformed staged image to the user so that the user can view the processed cell analysis result data along with the template data in the same image. Thus, the user can visually confirm how the processed cell analysis result data is compared with the template data. This data is provided in addition to a score representing the measured similarity calculated by the system. In addition, the system may display the processed target sample data and template data for user confirmation. The processed target sample data may be displayed using display attributes that are different from the display attributes of the template data. For example, processed target sample data can be displayed, for example, in one color, texture, brightness, etc., while template data is displayed, for example, in a different color, texture, brightness, etc. As a result, the user can more easily distinguish between the two data sets. As an alternative, the user may turn on or off the display of the template.

  Alternatively, the display may be presented to the user, including original histogram data of cell analysis result data.

(display)
6 (a) and 6 (b) show exemplary displays provided to the user upon completion of the process described in FIG. 6A to 6B show template matching between the processed cell analysis result data and the template data. The pink shaded area indicates the position of the normal sample template, and the black dots represent processed cell analysis result data. FIG. 6 (a) shows a sample with a high matching score indicating that it is a normal sample. FIG. 6 (b) shows a sample with a low matching score indicating that it is an abnormal sample.

(Template data)
The template data used in the system 100 can be standard template data or can be customized by a physician. The template data may represent a normal and healthy sample or an abnormal sample. Standard template data is data that can be intentionally selected and processed using thousands of samples. Furthermore, by using this template, noise and bias can be removed, ultimately becoming more objective than that aggregated by any user. In addition, there may be different template data for each variable in one analysis, providing multivariate or multiparameter analysis. In addition, there may be different templates representing one-dimensional, two-dimensional, or three-dimensional formats to provide more data for comparison with target sample data. In order to provide template data according to the present invention, it is possible to obtain multiple specific disease templates from the current patient sample or target sample.

  As discussed herein, by providing template data, the target sample is compared to the template, for example to identify anomalies in the target sample based on a special graphic pattern that can be displayed on the display. obtain. These abnormalities are associated with hemoglobin such as chronic lymphocytic leukemia (CLL), acute lymphocytic leukemia (ALL), chronic myelogenous leukemia (CML), acute myeloid leukemia (AML), eg thalassemia. It may include abnormalities, sickle cell development, and the like.

(Conclusion)
Modifications and adaptations of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. The above description of implementations of the invention has been presented for purposes of illustration and description. It is not exhaustive and does not limit the invention to the precise form disclosed. Modifications and variations are possible in light of the above teachings or may be derived from practice of the invention. For example, the described implementation includes software, however, systems and methods consistent with the present invention may be implemented as a combination of hardware and software, or as hardware alone.

  In addition, although aspects of the present invention have been described as being stored in memory, those skilled in the art will also recognize that these aspects are also secondary to, for example, a hard disk, floppy disk, or CD-ROM secondary. It is understood that other types of computer readable media may be stored such as storage devices, the Internet or other propagation media, or other types of RAM or ROM.

FIG. 1 is an exemplary schematic diagram of a system environment in which systems and methods may be implemented consistent with the principles of some embodiments of the invention. FIG. 2 is an exemplary schematic diagram of the major components of a computer consistent with the principles of some embodiments of the invention. FIG. 3 is an exemplary schematic diagram of server components consistent with the principles of some embodiments of the present invention. FIG. 4 shows an exemplary flow diagram of the steps performed by a computer consistent with the principles of some embodiments of the invention. FIGS. 5 (a) and 5 (b) are illustrations comparing similarity measurements between an original histogram image and a transformed stepped image consistent with the principles of some embodiments of the present invention. A typical display. 6 (a) and 6 (b) illustrate exemplary displays provided to the user consistent with the principles of some embodiments of the present invention. FIG. 7 shows an exemplary flow diagram illustrating steps performed by a computer consistent with the principles of some embodiments of the invention.

Claims (24)

A method of processing image data representing cell analysis result data,
Accessing image data;
Generating a density histogram based on the image data;
Converting the density histogram into a stepped image;
Performing a normalized cross-correlation between the stepped image and a reference image to measure similarity;
Determining anomalies based on the measured similarity;
Including the method.   The method of claim 1, wherein generating the density histogram includes generating a density compensation function to homogenize the image data.   The method of claim 2, wherein transforming the density histogram comprises thresholding the homogenized image data.   The method of claim 1, wherein generating the density histogram comprises low pass filtering the accessed image data.   The method of claim 1, wherein the anomaly is determined based on the measured similarity below a predetermined threshold.   The method of claim 1, wherein the reference image is accessed from a database including a plurality of reference images representing a plurality of abnormal patterns.   The method of claim 1, wherein a flagging process is performed to identify a pattern in the staged image based on the staged image, pattern change prediction, and the measured similarity.   The staged image and the reference image are displayed on a display, the staged image is displayed using a first display attribute, and the reference image is displayed using a second display attribute. The method according to claim 1. A system for processing image data representing cell analysis result data, the system comprising:
Memory for storing instructions;
A processor executing the instructions to implement the method;
The method comprises:
Accessing image data;
Generating a density histogram based on the image data;
Converting the density histogram into a stepped image;
Performing a normalized cross-correlation between the stepped image and a reference image to measure similarity;
Determining anomalies based on the measured similarity;
Comprising
system.   The system of claim 9, wherein generating the density histogram includes generating a density compensation function to homogenize the image data.   The system of claim 10, wherein transforming the density histogram includes thresholding the homogenized image data.   The system of claim 9, wherein generating the density histogram includes low pass filtering the accessed image data.   The system of claim 9, wherein the anomaly is determined based on the measured similarity below a predetermined threshold.   The system of claim 9, wherein the reference image is accessed from a database including a plurality of reference images representing a plurality of abnormal patterns.   The system of claim 9, wherein a flagging process is performed to identify a pattern of the staged image based on the staged image, pattern change prediction, and the measured similarity.   The staged image and the reference image are displayed on a display, the staged image is displayed using a first display attribute, and the reference image is displayed using a second display attribute. The system according to claim 9. A computer readable medium for storing instructions to be executed by a processor for performing a method of processing image data representing data of a cell analysis result, the method comprising:
Accessing image data;
Generating a density histogram based on the image data;
Converting the density histogram into a stepped image;
Performing a normalized cross-correlation between the stepped image and a reference image to measure similarity;
Determining anomalies based on the measured similarity;
Including
Computer readable medium.   The computer-readable medium of claim 17, wherein generating the density histogram includes generating a density compensation function to homogenize the image data.   The computer readable medium of claim 18, wherein transforming the density histogram comprises thresholding the homogenized image data.   The method of claim 15, wherein generating the density histogram comprises low pass filtering the accessed image data.   The computer-readable medium of claim 17, wherein the anomaly is determined based on the measured similarity below a predetermined threshold.   The computer-readable medium of claim 17, wherein the reference image is accessed from a database that includes a plurality of reference images representing a plurality of abnormal patterns.   The computer-readable medium of claim 17, wherein a flagging process is performed to identify a pattern of the staged image based on the staged image, pattern change prediction, and the measured similarity.   The staged image and the reference image are displayed on a display, the staged image is displayed using a first display attribute, and the reference image is displayed using a second display attribute. The computer-readable medium of claim 17.

Images